March 2009 - jboss-dev-forums - Jboss List Archives

[Design of Messaging on JBoss (Messaging/JBoss)] - Re: Another DeadLock, something around InVM

by clebert.suconic＠jboss.com

a pastebin would be better on this case: http://pastebin.com/f2f0de30d View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4215012#4215012 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4215012

17 years, 1 month

1
0
0 / 0

[Design the new POJO MicroContainer] - Re: AbstractVFSParsingDeployer doesn't kick in upon server r

by heiko.braun＠jboss.com

Thanks. Do you know the JIRA number so I can monitor it? View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4215011#4215011 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4215011

17 years, 1 month

1
0
0 / 0

[Design of Messaging on JBoss (Messaging/JBoss)] - Re: Another DeadLock, something around InVM

by clebert.suconic＠jboss.com

I was trying to replicate this on a test. I came up with a test, but it hangs with a different thread dump. it would probably be a good test. It starts a bunch of threads, all of them missing pings, and all of them closing the session. I'm playing with a random wait between the createSession and close: Thread.sleep(PING_INTERVAL * (RandomUtil.randomPositiveInt() % 3)); So, I could eventually have session.close and failure detection being called at the same time. WDYT ? : | /* | * Test the client triggering failure due to no pong received in time | */ | public void testMultiThreadOpenAndCloses() throws Exception | { | final TransportConfiguration transportConfig = new TransportConfiguration("org.jboss.messaging.integration.transports.netty.NettyConnectorFactory"); | | Interceptor noPongInterceptor = new Interceptor() | { | public boolean intercept(final Packet packet, final RemotingConnection conn) throws MessagingException | { | log.info("In interceptor, packet is " + packet.getType()); | if (packet.getType() == PacketImpl.PING) | { | log.info("Ignoring Ping packet.. it will be dropped"); | return false; | } | else | { | return true; | } | } | }; | | messagingService.getServer().getRemotingService().addInterceptor(noPongInterceptor); | final int numberOfSessions = 30; | final int numberOfThreads = 8; | final CountDownLatch flagStart = new CountDownLatch(1); | final CountDownLatch flagAligned = new CountDownLatch(numberOfThreads); | | final ClientSessionFactory csf = new ClientSessionFactoryImpl(transportConfig, | null, | DEFAULT_CONNECTION_LOAD_BALANCING_POLICY_CLASS_NAME, | PING_INTERVAL, | (long)(PING_INTERVAL * 1.5), | DEFAULT_CALL_TIMEOUT, | DEFAULT_CONSUMER_WINDOW_SIZE, | DEFAULT_CONSUMER_MAX_RATE, | DEFAULT_SEND_WINDOW_SIZE, | DEFAULT_PRODUCER_MAX_RATE, | DEFAULT_MIN_LARGE_MESSAGE_SIZE, | DEFAULT_BLOCK_ON_ACKNOWLEDGE, | DEFAULT_BLOCK_ON_NON_PERSISTENT_SEND, | DEFAULT_BLOCK_ON_PERSISTENT_SEND, | DEFAULT_AUTO_GROUP, | DEFAULT_MAX_CONNECTIONS, | DEFAULT_PRE_ACKNOWLEDGE, | DEFAULT_ACK_BATCH_SIZE, | DEFAULT_RETRY_INTERVAL, | DEFAULT_RETRY_INTERVAL_MULTIPLIER, | DEFAULT_MAX_RETRIES_BEFORE_FAILOVER, | DEFAULT_MAX_RETRIES_AFTER_FAILOVER); | | class LocalThread extends Thread | { | Throwable failure; | | @Override | public void run() | { | try | { | // Start all at once to make concurrency worst | flagAligned.countDown(); | flagStart.await(); | for (int i = 0; i < numberOfSessions; i++) | { | | ClientSession session = csf.createSession(false, false, false); | | Thread.sleep(PING_INTERVAL * (RandomUtil.randomPositiveInt() % 3)); | | session.close(); | } | } | catch (Throwable e) | { | failure = e; | } | } | }; | | LocalThread threads[] = new LocalThread[numberOfThreads]; | | for (int i = 0; i < numberOfThreads; i++) | { | threads = new LocalThread(); | threads.start(); | } | | flagAligned.await(); | flagStart.countDown(); | | Throwable e = null; | for (LocalThread t : threads) | { | t.join(); | if (t.failure != null) | { | e = t.failure; | } | } | | if (e != null) | { | throw new Exception("Test Failed", e); | } | | } | | View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4215008#4215008 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4215008

17 years, 1 month

1
0
0 / 0

[Design of Messaging on JBoss (Messaging/JBoss)] - Another DeadLock, something around InVM

by clebert.suconic＠jboss.com

*maybe* related to JORAM hangs: | [junit] Found one Java-level deadlock: | [junit] ============================= | [junit] "Thread-11705": | [junit] waiting to lock monitor 0xa8838124 (object 0xb276fcd8, a java.lang.Object), | [junit] which is held by "Thread-11703" | [junit] "Thread-11703": | [junit] waiting to lock monitor 0xa88381e4 (object 0xb279d280, a org.jboss.messaging.core.remoting.impl.invm.InVMConnection), | [junit] which is held by "Thread-11705" | [junit] | [junit] Java stack information for the threads listed above: | [junit] =================================================== | [junit] "Thread-11705": | [junit] at org.jboss.messaging.core.client.impl.ConnectionManagerImpl.failConnection(ConnectionManagerImpl.java:867) | [junit] - waiting to lock <0xb276fcd8> (a java.lang.Object) | [junit] at org.jboss.messaging.core.client.impl.ConnectionManagerImpl.connectionDestroyed(ConnectionManagerImpl.java:197) | [junit] at org.jboss.messaging.core.remoting.impl.invm.InVMConnector$Listener.connectionDestroyed(InVMConnector.java:187) | [junit] at org.jboss.messaging.core.remoting.impl.invm.InVMConnection.close(InVMConnection.java:96) | [junit] - locked <0xb279d280> (a org.jboss.messaging.core.remoting.impl.invm.InVMConnection) | [junit] at org.jboss.messaging.core.remoting.impl.invm.InVMConnector.disconnect(InVMConnector.java:161) | [junit] at org.jboss.messaging.core.remoting.impl.invm.InVMAcceptor$Listener$1.run(InVMAcceptor.java:167) | [junit] "Thread-11703": | [junit] at org.jboss.messaging.core.remoting.impl.invm.InVMConnection.close(InVMConnection.java:86) | [junit] - waiting to lock <0xb279d280> (a org.jboss.messaging.core.remoting.impl.invm.InVMConnection) | [junit] at org.jboss.messaging.core.remoting.impl.RemotingConnectionImpl.internalClose(RemotingConnectionImpl.java:557) | [junit] at org.jboss.messaging.core.remoting.impl.RemotingConnectionImpl.fail(RemotingConnectionImpl.java:423) | [junit] at org.jboss.messaging.core.client.impl.ConnectionManagerImpl.failConnection(ConnectionManagerImpl.java:871) | [junit] - locked <0xb276fcd8> (a java.lang.Object) | [junit] at org.jboss.messaging.core.client.impl.ConnectionManagerImpl.connectionDestroyed(ConnectionManagerImpl.java:197) | [junit] at org.jboss.messaging.core.remoting.impl.invm.InVMConnector$Listener.connectionDestroyed(InVMConnector.java:187) | [junit] at org.jboss.messaging.core.remoting.impl.invm.InVMConnection.close(InVMConnection.java:96) | [junit] - locked <0xb2775b98> (a org.jboss.messaging.core.remoting.impl.invm.InVMConnection) | [junit] at org.jboss.messaging.core.remoting.impl.invm.InVMConnector.disconnect(InVMConnector.java:161) | [junit] at org.jboss.messaging.core.remoting.impl.invm.InVMAcceptor$Listener$1.run(InVMAcceptor.java:167) | View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4214990#4214990 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4214990

17 years, 1 month

1
0
0 / 0

[Design the new POJO MicroContainer] - Re: InterruptedException not being cleared?

by david.lloyd＠jboss.com

OK, I'll try to come up with one... View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4214988#4214988 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4214988

17 years, 1 month

1
0
0 / 0

[Design the new POJO MicroContainer] - Re: InterruptedException not being cleared?

by scott.stark＠jboss.org

I think that looks ok, but we really need some tests to validate the expected behaviors for the ways multiple cooperating class loading threads can be interrupted. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4214986#4214986 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4214986

17 years, 1 month

1
0
0 / 0

[Design the new POJO MicroContainer] - Re: InterruptedException not being cleared?

by david.lloyd＠jboss.com

There's another interrupt-eater down below in the same file, though this one I think was just a mistake of omission: | @@ -386,16 +394,17 @@ | } | else | { | releaseInNextTask = true; | } | } | catch (InterruptedException ignored) | { | + interrupted = true; | } | thread = loadClassThreads.get(classLoader); | } | } | finally | { | if (interrupted) | Thread.currentThread().interrupt(); | View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4214982#4214982 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4214982

17 years, 1 month

1
0
0 / 0

[Design of JBoss Profiler] - Jboss profiler for Solaris on intel

by ciroBarradov

Hi, I'm trying to compile the Jboss profiler library (so) in a machine that uses Solaris on Intel. I tried to generate the so file, with solaris code, but it didn't worked due to my machine isn't SPARC architecture. Is there any source code for this kind of machines (SOLARIS ON INTEL)? Thanks View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4214970#4214970 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4214970

17 years, 1 month

1
0
0 / 0

[Design the new POJO MicroContainer] - Re: InterruptedException not being cleared?

by david.lloyd＠jboss.com

This trivial patch gets clears the interrupt status, tolerates interrupts, and then restores the interrupt status in a finally block. Is there a reason to not do it this way? Otherwise I'll commit it, if you give the OK. Of course the forums will mangle this, but... | Index: classloader/src/main/java/org/jboss/classloader/spi/base/ClassLoaderManager.java | =================================================================== | --- classloader/src/main/java/org/jboss/classloader/spi/base/ClassLoaderManager.java (revision 85240) | +++ classloader/src/main/java/org/jboss/classloader/spi/base/ClassLoaderManager.java (working copy) | @@ -185,133 +185,141 @@ | * Process the next task | * | * @param thread the thread | * @param task the task | * @throws InterruptedException if it is interrupted | */ | private static void nextTask(Thread thread, ClassLoadingTask task) throws InterruptedException | { | - boolean trace = log.isTraceEnabled(); | - if (trace) | - log.trace("Next task thread=" + thread + " task=" + task); | - | - List<ThreadTask> taskList = loadTasksByThread.get(thread); | - synchronized (taskList) | + boolean intr = Thread.interrupted(); | + try | { | - // There may not be any ThreadTasks | - while (taskList.isEmpty() && task.getThreadTaskCount() != 0 ) | + boolean trace = log.isTraceEnabled(); | + if (trace) | + log.trace("Next task thread=" + thread + " task=" + task); | + | + List<ThreadTask> taskList = loadTasksByThread.get(thread); | + synchronized (taskList) | { | - /* There are no more tasks for the calling thread to execute, so the | - calling thread must wait until the task.threadTaskCount reaches 0 | - */ | - if (trace) | - log.trace("Begin nextTask(WAIT_ON_EVENT), task="+task); | - try | + // There may not be any ThreadTasks | + while (taskList.isEmpty() && task.getThreadTaskCount() != 0 ) | { | - task.waitOnEvent(); | - taskList.wait(); | + /* There are no more tasks for the calling thread to execute, so the | + calling thread must wait until the task.threadTaskCount reaches 0 | + */ | + if (trace) | + log.trace("Begin nextTask(WAIT_ON_EVENT), task="+task); | + try | + { | + task.waitOnEvent(); | + taskList.wait(); | + if (trace) | + log.trace("nextTask(WAIT_ON_EVENT), notified, task="+task); | + } | + catch(InterruptedException e) | + { | + if( trace ) | + log.trace("nextTask(WAIT_ON_EVENT), interrupted, task="+task, e); | + intr = true; | + } | } | - catch(InterruptedException e) | + | + if (trace) | + log.trace("Continue nextTask(" + taskList.size()+"), task="+task); | + | + // See if the task is complete | + if (task.getThreadTaskCount() == 0) | { | - if( trace ) | - log.trace("nextTask(WAIT_ON_EVENT), interrupted, task="+task, e); | - // Abort this task t | - throw e; | + task.finish(); | + log.trace("End nextTask(FINISHED), task="+task); | + return; | } | - if (trace) | - log.trace("nextTask(WAIT_ON_EVENT), notified, task="+task); | } | | + ThreadTask threadTask = taskList.remove(0); | + ClassLoadingTask loadTask = threadTask.getLoadTask(); | if (trace) | - log.trace("Continue nextTask(" + taskList.size()+"), task="+task); | + log.trace("Begin nextTask(" + taskList.size() + "), loadTask=" + loadTask); | | - // See if the task is complete | - if (task.getThreadTaskCount() == 0) | + try | { | - task.finish(); | - log.trace("End nextTask(FINISHED), task="+task); | - return; | - } | - } | - | - ThreadTask threadTask = taskList.remove(0); | - ClassLoadingTask loadTask = threadTask.getLoadTask(); | - if (trace) | - log.trace("Begin nextTask(" + taskList.size() + "), loadTask=" + loadTask); | - | - try | - { | - Thread taskThread = threadTask.getThread(); | - if (taskThread == null) | - { | - /* This is a task that has been reassigned back to the original | - requesting thread ClassLoadingTask, so a new ThreadTask must | - be scheduled. | - */ | - if (trace) | - log.trace("Rescheduling threadTask=" + threadTask); | - scheduleTask(loadTask, threadTask.getLoader(), true); | + Thread taskThread = threadTask.getThread(); | + if (taskThread == null) | + { | + /* This is a task that has been reassigned back to the original | + requesting thread ClassLoadingTask, so a new ThreadTask must | + be scheduled. | + */ | + if (trace) | + log.trace("Rescheduling threadTask=" + threadTask); | + scheduleTask(loadTask, threadTask.getLoader(), true); | + } | + else | + { | + if (trace) | + log.trace("Running threadTask=" + threadTask); | + threadTask.run(); | + } | } | - else | + catch (Throwable e) | { | if (trace) | - log.trace("Running threadTask=" + threadTask); | - threadTask.run(); | - } | - } | - catch (Throwable e) | - { | - if (trace) | - log.trace("Run failed with exception", e); | - boolean retry = e instanceof ClassCircularityError || e.getClass().equals(LinkageError.class); | - if (retry && loadTask.incrementNumCCE() < MAX_CCE) | - { | - /* Reschedule this task after all existing tasks to allow the | - current load tasks which are conflicting to complete. | - */ | - try | + log.trace("Run failed with exception", e); | + boolean retry = e instanceof ClassCircularityError || e.getClass().equals(LinkageError.class); | + if (retry && loadTask.incrementNumCCE() < MAX_CCE) | { | - // Reschedule and update the loadTask.threadTaskCount | - scheduleTask(loadTask, threadTask.getLoader(), true); | + /* Reschedule this task after all existing tasks to allow the | + current load tasks which are conflicting to complete. | + */ | + try | + { | + // Reschedule and update the loadTask.threadTaskCount | + scheduleTask(loadTask, threadTask.getLoader(), true); | + } | + catch (Throwable ex) | + { | + loadTask.setLoadError(ex); | + log.warn("Failed to reschedule task after CCE", ex); | + } | + if (trace) | + log.trace("Post CCE state, loadTask=" + loadTask); | } | - catch (Throwable ex) | + else | { | - loadTask.setLoadError(ex); | - log.warn("Failed to reschedule task after CCE", ex); | + loadTask.setLoadError(e); | } | - if (trace) | - log.trace("Post CCE state, loadTask=" + loadTask); | } | - else | + finally | { | - loadTask.setLoadError(e); | + // Release any lock on the classloader | + if (threadTask.isReleaseInNextTask()) | + threadTask.getClassLoader().unlock(false); | } | - } | - finally | - { | - // Release any lock on the classloader | - if (threadTask.isReleaseInNextTask()) | - threadTask.getClassLoader().unlock(false); | - } | | - // If the ThreadTasks are complete mark the ClassLoadingTask finished | - if (loadTask.getThreadTaskCount() == 0) | - { | - List<ThreadTask> loadTaskThreadTasks = loadTasksByThread.get(loadTask.getRequestingThread()); | - synchronized (loadTaskThreadTasks) | + // If the ThreadTasks are complete mark the ClassLoadingTask finished | + if (loadTask.getThreadTaskCount() == 0) | { | - if( trace ) | - log.trace("Notifying task of thread completion, loadTask:"+loadTask); | - task.finish(); | - loadTaskThreadTasks.notify(); | + List<ThreadTask> loadTaskThreadTasks = loadTasksByThread.get(loadTask.getRequestingThread()); | + synchronized (loadTaskThreadTasks) | + { | + if( trace ) | + log.trace("Notifying task of thread completion, loadTask:"+loadTask); | + task.finish(); | + loadTaskThreadTasks.notify(); | + } | } | + if (trace) | + log.trace("End nextTask(" + taskList.size()+ "), loadTask=" + loadTask); | + } | + finally | + { | + if (intr) | + Thread.currentThread().interrupt(); | } | - if (trace) | - log.trace("End nextTask(" + taskList.size()+ "), loadTask=" + loadTask); | } | | /** | * Invoked to create a ThreadTask to assign a thread to the task of | * loading the class of ClassLoadingTask. | * | * @param task the classloading task | * @param loader the loader | View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4214969#4214969 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4214969

17 years, 1 month

1
0
0 / 0

[Design the new POJO MicroContainer] - Re: InterruptedException not being cleared?

by scott.stark＠jboss.org

Yes, that is problem, what to do with the exception we saw. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4214965#4214965 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4214965

17 years, 1 month

1
0
0 / 0

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

jboss-dev-forums March 2009